A Digital Imax Camera

[Chris D Walker]

A while ago there was talk about Imax making a digital cinema camera. There were no specs drawn up and I haven't heard anything more since, so the question becomes what would make a good Imax digital camera? Can we speculate what the Imax boffins could realistically build? I'll take a stab at it.

 

First, the sensor would need to be 70mm wide and have a 1.44:1 ratio or close. Maybe like the Alexa and higher-end film scanners, which oversample 3K to 2K, the sensor could oversample 6K to 4K. Without doing much math, the sheer size of each photosite would give the camera an extra two stops of sensitivity over S35 chip designs; a conservative rating of 2000ASA native.

 

I think an oversampled 4K image printed onto 70mm Imax film or screening from a 4K digital projector would be more than enough resolution. It would be a law of diminishing returns for the viewer as the resolution goes beyond 4K in terms of noticing a significant difference.

 

With regard to colour resolution and a compression scheme (if any), I'm not so sure. The number of people and venues that could fully utilise such a camera would be very small; much the same as it is now with film Imax cameras.

 

Any differing opinions? Different specs? Is it worth building such a camera?

 

Thanks to those who read and reply.

[Adrian Sierkowski]

http://www.visionresearch.com/Products/High-Speed-Cameras/Phantom-65/

[David Mullen ASC]

If a digital camera is going to replace the 15-perf 65mm camera it has to be more than 4K. Even Red says that their tests of Super-35 film show that it resolves 3.2K or so. The 15-perf 65mm frame is 2.9X wider than that. Simple math says that if your piece of film is 2.9X wider, it would be resolving over 9K. Even if you count the limitations of large-format optics and their lower MTF, and the lower MTF of film in general, it's not like it's going to drop down to 4K if 35mm measures at 3.2K.

 

So a 4K camera is not going to be a true replacement, resolution-wise, to a 15-perf 65mm camera.

 

Seems to me that IMAX should be aiming for a 6K system, or even 8K. Already you have some people shooting 4K for ordinary movies, and some theaters are getting 4K projectors, and once those 4K movies start going out in 4K DCP's, then exactly how is IMAX going to promise a radically better moviegoing experience if they are also centered around a 4K system? Why not at least simply double what everyone else is getting, considering that the screen is twice as large anyway?

 

For IMAX 3D work, probably a decent 4K camera would work because it's hard to judge resolution when you've got the 3D effect screwing your eyeballs around.

 

IMAX was always about the promise of knocking your socks out with the degree of fine detail on a giant screen, not about finding "good enough" solutions that you can just get away with if the audience isn't paying close attention.

 

We're talking about IMAX here, Chris, the biggest movie format ever invented -- dream bigger!

[Chris D Walker]

You make a good point. Although, remember that I'm the person who feels 2K is good enough and that anything above 4K is overkill, especially when you're in a dark theatre more than 20ft. away from the screen. That's what bothers me about those demos NHK present with 8K cameras and projectors; people stand directly in front of the screen to notice the difference because the average persons' eyes can't from a viewing distance.

 

Is resolution the only difference that Imax provides with its massive negative? How many of us could actually see the difference between a 4K and an 8K projection if we were blind-tested? I doubt I could.

 

But this is the point of the discussion; to ask what would it take to produce a digital Imax camera in your opinion. I feel greater sensitivity and dynamic range would be of greater benefit than a marginal visual improvement from 4K to 8K that would eat four times as much data and bandwidth. But that's me. I enjoy a constructive disagreement.

[Ravi Kiran]

Is resolution the only difference that Imax provides with its massive negative? How many of us could actually see the difference between a 4K and an 8K projection if we were blind-tested? I doubt I could.

 

I can't imagine that on a true IMAX-sized screen the extra resolution wouldn't make a difference, but I doubt we'd be able to tell the difference between 4K and 8K on a regular screen. But that's just my untested opinion...

[Brian Rose]

Imax for me has always been about absurdity. I mean, 65mm film is big and costly and unwieldy as 5-perf vertical pulldown. So what do they do? The say, "Hell, let's flip the sucker and shoot 15 perf!" They design cameras that can go to Everest, into space, or 12,000 feet beneath the sea. And then they throw it up on the biggest F'in screen you've ever seen.

 

I'm willing to be open minded about digital, so long as they bring that same "to hell with common sense" approach. They need to come up with an ABSURD digital camera. I say forget 8K. Go for 12K, and shoot at 48 FPS or even 60 FPS, to finally put to bed the old strobing issues that have plagued 24 FPS Imax. Go all out, Imax!

[Antti Näyhä]

4k is definitely not enough for a real IMAX screen (not talking about those sad multiplex installations). On Transformers 2, they did a 4k DI for the IMAX shots, and it was quite obvious that the resolving power was compromised compared to contact-printed 15/65mm IMAX.

[Bruce Greene]

A while ago there was talk about Imax making a digital cinema camera. There were no specs drawn up and I haven't heard anything more since, so the question becomes what would make a good Imax digital camera? Can we speculate what the Imax boffins could realistically build? I'll take a stab at it.

 

First, the sensor would need to be 70mm wide and have a 1.44:1 ratio or close. Maybe like the Alexa and higher-end film scanners, which oversample 3K to 2K, the sensor could oversample 6K to 4K. Without doing much math, the sheer size of each photosite would give the camera an extra two stops of sensitivity over S35 chip designs; a conservative rating of 2000ASA native.

 

I think an oversampled 4K image printed onto 70mm Imax film or screening from a 4K digital projector would be more than enough resolution. It would be a law of diminishing returns for the viewer as the resolution goes beyond 4K in terms of noticing a significant difference.

 

With regard to colour resolution and a compression scheme (if any), I'm not so sure. The number of people and venues that could fully utilise such a camera would be very small; much the same as it is now with film Imax cameras.

 

Any differing opinions? Different specs? Is it worth building such a camera?

 

Thanks to those who read and reply.

 

By coincidence I did a little comparison the other night.

 

I happened to shoot a still photo in two formats:

 

1. 6x9cm color negative film (ISO 160) or about IMAX size

2. Full frame Canon 5d Mk1 13mp

 

Both versions were shot on a tripod and the film was scanned on a dedicated filmscanner at 8k

 

The result was that the film was a little bit more detailed, but with grain often obscuring the fine detail. I judged the results from each camera pretty close in image quality.

 

So my speculation is that a vista vision sized chip (still camera, full 35mm frame size) with about 26mp (4.5k Bayer array) will equal the current IMAX film detail with far less grain/noise

[Adrian Sierkowski]

I don't know if I would call that conclusive.. I think you'd need motion samples of both projected in motion, for one, and for two to be use to use and not which film you're going with. There are subtle differences, to my eye, between motion picture stock, and stills stock which I feel will magnify as the image grows larger...

[K Borowski]

An IMAX frame is, very conservatively, 8K (8192x5688 46.6MP). If you were shooting 5201, or F64D, the resolution would probably far exceed that.

 

To say that it is only 12MP is laughable. That would make a frame of anamorphic 12/~9.72=1.23MP per frame. Maybe the scan sucks, the camera' flange focal distance is off, or you shot P3200 film that expired in 2003 by mistake.

[Mitch Gross]

There is a Digital IMAX camera, or at least the prototype. The images from it are stunning. Wish I could tell you more, but I suggest you go see "Born To Be Wild" when it hits IMAX 3D screens in March. Here's the trailer:

 

http://www.youtube.com/watch?v=ieqGOszVFzI

 

A lot of the suppositions made in this thread are based on incomplete knowledge. I can tell you that this footage blows away any digital motion picture imaging you've seen elsewhere. It's very impressive.

[Bruce Greene]

An IMAX frame is, very conservatively, 8K (8192x5688 46.6MP). If you were shooting 5201, or F64D, the resolution would probably far exceed that.

 

To say that it is only 12MP is laughable. That would make a frame of anamorphic 12/~9.72=1.23MP per frame. Maybe the scan sucks, the camera' flange focal distance is off, or you shot P3200 film that expired in 2003 by mistake.

 

Hi Karl,

 

I stand by my observation :)

 

Feel free to conduct your own test and report back. I'd be interested.

 

Assumptions made from one's imagination or web posts are not evidence, even if you're smart...

[Bruce Greene]

I don't know if I would call that conclusive.. I think you'd need motion samples of both projected in motion, for one, and for two to be use to use and not which film you're going with. There are subtle differences, to my eye, between motion picture stock, and stills stock which I feel will magnify as the image grows larger...

 

I would expect that in motion, the film grain will be less visible, but the detail rendered might be reduced by registration errors. Maybe a wash.

[Adrian Sierkowski]

True true, but you'd get the same "registration issues," if you were to print and project your digital footage, so I still think the comparison would be important, you know. Hell, even when I freeze frame some film I've shot, I hate it! The grain urgh! of course, in motion it's a whole other story as I no longer really "notice," and it appears, to me at least, sharper in movement.

[Anton Papich]

 

Both versions were shot on a tripod and the film was scanned on a dedicated filmscanner at 8k

 

The result was that the film was a little bit more detailed, but with grain often obscuring the fine detail. I judged the results from each camera pretty close in image quality.

 

So my speculation is that a vista vision sized chip (still camera, full 35mm frame size) with about 26mp (4.5k Bayer array) will equal the current IMAX film detail with far less grain/noise

 

And what scanner would that be?

[John Sprung]

... what would make a good Imax digital camera?

 

We could dust off the Lockheed-Martin "Blue Herring" camera of the 1990's. It used Imax-size chips. The amazing thing was, it was a three chip design, 12 mega-photosites on each chip. It made wonderful pictures, but the prototype was large, between the size of a Technicolor blimp and a Mini-Cooper.

 

 

 

 

-- J.S.

[David Mullen ASC]

I don't get it -- 35mm film is around 3.2K but a replacement for 15-perf 65mm only needs to be 4K??? On what planet??? So all these years, regular 35mm film was nearly IMAX quality and we never realized it?

 

If you really believe that a 15-perf 65mm negative only resolves about 4K, you'd have to believe that 35mm negative barely resolves 1.3K (less than HD) because the 15-perf 65mm negative is nearly 3X wider.

 

It would be an absolute f---ing TRAGEDY if 15-perf 65mm photography was replaced by 4K digital photography.

 

IMAX, as that format that blew our socks off on a 50' screen, would disappear as we know it, to be replaced by something "good enough" that was not dramatically different or better than, let's say, 5-perf 65mm.

 

It would be the final triumph of the bean-counters, destroying the ultimate film format ever invented.

 

Anyone with a Red One camera can shoot 4K right now, so does that mean every Red One owner out there is actually shooting IMAX-quality footage? Gee, for some reason when I saw "Fair Game" I didn't realize it was shot in IMAX.

 

So 3.2K is 35mm quality and 4K is 15-perf 65mm quality? I didn't realize there was such a tiny difference between a 24mm x 18mm negative and a 70mm x 52mm negative, I guess it's more like the jump when you go from 16mm to Super-16mm.

 

It's one thing to say that 4K works fine for IMAX blow-ups, it's another to say that it equals 15-perf 65mm and we can get rid of the old format and not losing anything in the process.

 

4K replacing 15-perf 65mm is not a dream, it's the killer of dreams. It's like you're taking one of my dogs out into the yard and shooting it in the head.

[K Borowski]

It would be an absolute f---ing TRAGEDY if 15-perf 65mm photography was replaced by 4K digital photography.

 

[. . . ]

 

It would be the final triumph of the bean-counters, destroying the ultimate film format ever invented.

 

[. . .]

 

4K replacing 15-perf 65mm is not a dream, it's the killer of dreams. It's like you're taking one of my dogs out into the yard and shooting it in the head.

 

No, what IMAX didn't realize is that a digital revolution is sweeping the nation. Anything film must be replaced as quickly as possible. David, you've said it far more eloquently than the response I wrote but deleted. Once I realized that there is no point in arguing anymore (my experiences with 70mm 15-perforation are anything but internet speculation), I deleted it.

 

Someone who wants to believe that 12 MP (even 26, let's say 26) is better than 70mm, 15 perforation IMAX prints that are contact-printed off the OCN is going to believe what they want to believe no matter how well backed-up your argument may be.

 

 

 

Your last two sentences I've quoted, I completely agree with. Yet, your first quoted sentence would be a better fate than is befalling many IMAX theatres right now: 15-perf. 70mm projectors with astonishing optics, vacuum gates, that take a chain of film 30 feet up in the air, project it, and take it back onto a platter, often in a humidity controlled environment that the public can see through glass windows as they wander in - are being replaced by a pair of 2K projectors with at best half the resolution of a single 4K machine.

 

It is a tragedy, but we live in a world full of tragedies, not Hollywood endings. . .

[David Mullen ASC]

IMAX was always special, that really over-the-top format, a HUGE leap over what Hollywood used for regular movies. 4K is not over-the-top, it's not light years better than what a typical movie is shot on -- and some of them are shooting in 4K. There is no "wow" factor by saying "instead of shooting our movie in regular 4K, we're going to shoot our movie... in IMAX 4K! It will be so much better than your 4K movie because our 4K picture will be square. Once you see 4K on the big screen you'll never go back to, errr... 4K." You see how underwhelming it all sounds? Instead of shooting something 3X bigger and better than even regular 65mm, we're talking about shooting in something only marginally better than what anyone can shoot.

 

I guess that's great for a lot of people's egos out there, that your typical film student with a Red One is shooting near IMAX quality work and he doesn't yet realize it. But once people realize that IMAX has gone 4K and become as common as dirt, it will lose its allure. The name will mean nothing, it won't mean that you're seeing something that is an incredible leap far beyond what ordinary movies can achieve.

[Vincent Sweeney]

Hi Karl,

 

I stand by my observation :)

 

Feel free to conduct your own test and report back. I'd be interested.

 

Assumptions made from one's imagination or web posts are not evidence, even if you're smart...

 

I'm confused here Bruce, you do understand that motion film has nothing at all to do with still photos, right? Random grain being repeated 24 times a second fills in a detail gap, after gap, after gap, giving even the small 16mm frame great perceivable detail when in motion. Repetition of random grain and the resulting images of a 15perf 65mm frame seems hard to even comprehend and certainly has to be close to "12K" in digital terms, although I hate to use that word when talking IMAX.

 

Not sure about the registration errors you are talking about. I'm an avid IMAX viewer and haven't seen this. In fact all I see are images that are more immersing than 3D because of its resolution. Their system is pretty damn well designed I think.

 

Wait, what is Mitch referring to? Speak! ;)

[Tim Tyler]

There is a Digital IMAX camera, or at least the prototype. The images from it are stunning. Wish I could tell you more, but I suggest you go see "Born To Be Wild" when it hits IMAX 3D screens in March.

 

I saw this trailer last night in 3D IMAX before Tron:Legacy. It was pretty amazing to see. There was amazing clarity and a real perspective that made it stand out among the other 3D trailers (although some of those were 3D added in post movies.) I did feel that some of the shots had odd interocular distances though.

[David Mullen ASC]

I notice that supporters of this 4K IMAX idea can't explain why, if 35mm is 3.2K-ish, that 15-perf 65mm IMAX is somehow only 4K. How can such a HUGE increase in negative size yield only an improvement from 3.2K to 4K?

[Bruce Greene]

If a digital camera is going to replace the 15-perf 65mm camera it has to be more than 4K. Even Red says that their tests of Super-35 film show that it resolves 3.2K or so. The 15-perf 65mm frame is 2.9X wider than that. Simple math says that if your piece of film is 2.9X wider, it would be resolving over 9K. Even if you count the limitations of large-format optics and their lower MTF, and the lower MTF of film in general, it's not like it's going to drop down to 4K if 35mm measures at 3.2K.

 

So a 4K camera is not going to be a true replacement, resolution-wise, to a 15-perf 65mm camera.

 

Seems to me that IMAX should be aiming for a 6K system, or even 8K. Already you have some people shooting 4K for ordinary movies, and some theaters are getting 4K projectors, and once those 4K movies start going out in 4K DCP's, then exactly how is IMAX going to promise a radically better moviegoing experience if they are also centered around a 4K system? Why not at least simply double what everyone else is getting, considering that the screen is twice as large anyway?

 

For IMAX 3D work, probably a decent 4K camera would work because it's hard to judge resolution when you've got the 3D effect screwing your eyeballs around.

 

IMAX was always about the promise of knocking your socks out with the degree of fine detail on a giant screen, not about finding "good enough" solutions that you can just get away with if the audience isn't paying close attention.

 

We're talking about IMAX here, Chris, the biggest movie format ever invented -- dream bigger!

 

This film vs. digital origination seems to be somewhat of an emotional subject for some reason...

 

My observation was not based on math, but by looking at a piece of scanned IMAX sized film and comparing it to a VistaVision sized 35mm chip original. My curiosity was really to determine if I should keep my 6x9cm still cameras and scanner or just switch to digital for still photography.

 

I was somewhat surprised by the results myself, and I understand that a movie will show less grain when displaying 24fps. But, detail is detail and it was not the difference I would have expected to find, but it is what I observed.

 

Both formats were well focused, and the scan was very well focused by virtue of the detail of the film grain.

 

I think the main point is that a lot more detail is lost shooting on color negative film, and then scanning vs. digital origination of the original scene.

 

Please also read my original post. I never suggested that a RED camera at 4k is equal to IMAX film. I did suggest that 4k digital of VistaVision size chip will come close to IMAX film scanned, even at an 8k scan.

 

I also suspected that I would get some pointed response to my post, but I was just reporting my observation, knowing that there would be some disagreement. It's a tough crowd here on the forum:)

 

Of course, a digital replacement for IMAX should be a substantial improvement over the current system, not just approximate what already works really well.

[David Mullen ASC]

This film vs. digital origination seems to be somewhat of an emotional subject for some reason...

 

My observation was not based on math, but by looking at a piece of scanned IMAX sized film and comparing it to a VistaVision sized 35mm chip original. My curiosity was really to determine if I should keep my 6x9cm still cameras and scanner or just switch to digital for still photography.

 

I was somewhat surprised by the results myself, and I understand that a movie will show less grain when displaying 24fps. But, detail is detail and it was not the difference I would have expected to find, but it is what I observed.

 

Both formats were well focused, and the scan was very well focused by virtue of the detail of the film grain.

 

I think the main point is that a lot more detail is lost shooting on color negative film, and then scanning vs. digital origination of the original scene.

 

Please also read my original post. I never suggested that a RED camera at 4k is equal to IMAX film. I did suggest that 4k digital of VistaVision size chip will come close to IMAX film scanned, even at an 8k scan.

 

So a 4K sensor that is 24mm x 13mm, let's say, has less resolution than a 4K sensor that is 36mm x 24mm? I mean, it's certainly possible because the larger sensor would not need a lens with as high an MTF as the smaller sensor, but if a sensor is 4K RAW, for example, it can't ever end up with more than 80% or so of that pixel resolution in measurable resolution once debayered into RGB -- you can't shoot with a 4K sensor and end up with 6K of measurable resolution, for example. And the improvement in MTF of the larger sensor isn't going to overcome that limitation nor cause the resolution to effectively double.

 

To some extent when people compare a high megapixel digital still to a film scan and noticing that the digital image seems superior or similar in resolution but with less grain, it's partly due to the nature of digital images, the higher MTF and edge sharpness. Film is "detailed yet softer" and digital is "less detailed but sharper". This is one reason why 1080P HD images, though in theory being around 1.6K or so in measurable line resolution compared to 35mm at 3.2K-ish, seem sharper in terms of fine detail in close-ups, because edges and lines and other visual clues as to sharpness are not softened as film does, they have a crispness to them, even if no edge enhancement is applied (and to some small degree, aliasing has some ability to fool the eye that they are seeing more detail.)

 

But after the initial excitement about 1080P HD in the early 2000's, with people like Lucas and Cameron saying that it was not only just as good as 35mm, it was more like 70mm in terms of sharpness, we know now from a decade of looking at 1080P images on the big screen that the actual resolution is somewhat sub-35mm, it just LOOKS sharp - I mean, it IS sharp, but it's not as detailed. Wide shots of "soft" information (like nature) without hard edges to define shapes, tend to show how 1080P sort of craps out compared to 35mm.

 

Also remember that with a piece of film, information (grains) exist in a different spot on each frame, so the image seems more detailed as it is being projected due to some mental averaging of detail as the frames flash by, the higher the frame rate of the projector, the better. But with digital, the information exists in the same spot on each frame, so there is less visible improvement in resolution with faster projection.

 

So I think comparing still frames, though useful, does not give you the complete picture so to speak as to the differences between film and digital. Still frames of movie film have always looked softer than they do once projected in a series of frames.

 

I saw the IMAX trailer of "Born to be Wild" as well, before "Tron: Legacy" -- it looked good (the digital shots were mixed in, it wasn't all digitally-shot) but the problem with trying to judge resolution with a 3D movie is that your eyes are dealing with two images, so the 3D effect takes over as the primary effect and it's hard to concentrate on fine detail being resolved across the frame. Plus I have no frame of reference and I don't know how the trailer was made, maybe it was all done at 4K.

 

What I'd really need to see are line resolution charts shot on 15-perf 65mm negative and scanned at 8K or higher and compare that to tests of the 4K IMAX camera pointed at charts, plus see the 4K IMAX charts recorded out to 15-perf 65mm and projected side-by-side with the 15-perf 65mm tests contact printed to see what loss occurs over projection.

 

Look, from a practical standpoint, it's understandable, this need for IMAX to embrace a camera that is much easier and cheaper to use than a 15-perf 65mm camera. But I'm not talking about what's practical, I'm only talking about the IDEAL that IMAX represents. And I'm sure that there is some improvement from using a larger 4K sensor over a smaller 4K sensor in terms of noise and MTF (though lower depth of field may also contribute to the feeling of increased sharpness). And I'm sure it blows up to IMAX fine -- after all, "Tron" didn't look too bad and that was 1080P F35 photography blown-up to IMAX. But ultimately 4K has to be a temporary stopgap measure -- for IMAX to maintain this ideal of offering radically superior image quality over regular 35mm and 65mm photography, it has to think really, really big.

 

Probably the closest thing on the book is the proposed 6K FF35 "Monstro" sensor from Red. THAT I could see as being IMAX-like if projected digitally in 6K.

[Bruce Greene]

Hi David,

 

Thanks for your thoughtful and well written post!

 

My little observation pretty much agrees entirely with your post, especially the conclusion about the "Monstro" sensor.

 

And yes the "aliasing" was quite evident in my observation, but in the end, it was about the final impression of the image. I mean, my film image probably had twice the resolution of my digital version, but the film certainly didn't look twice as detailed, if that makes sense. In some parts of the image I could "see" more detail in the digital image due to high MTF and aliasing. It was just very interesting to me that such a large piece of color negative film looked so similar to such a low resolution digital capture of the same scene. Unexpected, but observable.

 

My "pot stirring" point, I guess, was not to let our preconceptions and romanticism about film capture lead to erroneous conclusions about it's superiority. I think this is evident in the responses that questioned my scanner or my focus or my flange depth...Sometimes I think we turn photography into religious positions, but that's a pot to stir at another time:)